Location of the experiment

This research was carried out in the nursery and in the phytopathology laboratory of the experimental station Juan Bernito of ict, located in the district of La Banda de Shilcayo, province and region of San Martin, Peru, at 76°00´18" W and 06°00´28" S, at an altitude of 315 m.a.s.l. The temperature, relative humidity and precipitation were 25 ° C, 73 % and 1,188 mm, respectively.

Plant material

For the isolation and identification of the causal agent of the sudden death of T. cacao, samples of affected tissues with death symptoms and plants with "sudden death" typically affected by Verticillium were collected in January 2010 on a two-year-old CCN-51 cacao plantation of the Atusparia sector, in the district and province of Tocache in the San Martin region of Peru. In this plantation, 5 % of the plants exhibited symptoms of leaf yellowing, widespread wilting and plant death. When performing transverse cuts of the stem, internal vascular occlusion was observed as points, meanwhile in the longitudinal cuts, brown striae were observed in the vascular system. Samples were taken to the ict Phytopathology Laboratory for their isolation and identification.

Isolation and identification

Isolation of the causative agent was obtained from segmented tissues of the stem, root and neck of diseased plants. The segments were washed with water and disinfested with 1 % sodium hypochlorite for 10 minutes to remove surface microorganisms; in addition, they were washed with sterile distilled water and placed on paper towels for drying. From each segment, 1 cm portions were cut and planted in Petri dishes containing potato dextrose agar medium with oxytetracycline (pdao), and incubated for 7 days at 25 °C (^{Resende, Flood, & Cooper, 1995}). The plates were subjected to artificial illumination at a room temperature of 28 °C ± 0.57 °C and a relative humidity of 70 % ±} 1.71 % for 10 days ±} 1 to favor fungus sporulation. The taxonomic identification of the fungus was made by observing the macroscopic morphological characteristics of the colony (appearance and coloration) as well as the microscopic morphological characteristics (type and size of vegetative and reproductive structures). The keys published by Barnett and Hunter (¹⁹⁹⁸), Barron (¹⁹⁶⁸) and Watanabe (²⁰¹⁰) were used for identification of the pathogen.

Further, the identification of the isolates was performed using Sanger's sequencing. For this purpose, the isolates were cultivated in flasks containing 50 mL of potato dextrose broth (pdb, 24 g/L) at 25 °C until the mycelium colonized the surface of the broth. The mycelium was washed, filtered and stored at -20 °C until dna extraction was carried out. The mycelium was triturated in liquid nitrogen and the dna was extracted using the DNeasy PowerPlant Pro kit (Qiagen, Germany), following the manufacturer's instructions. Partitions ITS1/ITS4 (ITS1: 5'-tccgtaggtgaacctgcgg-3'; ITS4: 5'-tcctccgcttattgatatgc-3') (^{White, Bruns, Lee, & Taylor, 1990}) were used to amplify dna.

The pcr mixture was prepared using Master Mix 2x pcr (Promega, U.S.A.), in a final volume of 25 μL containing Milli-Q water, 1x MasterMix pcr, 0.5 μM of each 5' to 3' partitions (forward or direct), 3' to 5' (reverse) partitions and 2 μL of cDNA. The pcr was performed on a Techne TC-412 thermal thermocycler (Techne). The conditions employed were 94 °C for 2 minutes, followed by 35 cycles of 94 °C for 10 seconds, 54 °C for 30 seconds and 72 °C for 1 minute, with a final elongation phase of 72 °C for 10 minutes. Amplicons were monitored by agarose gel electrophoresis at 1 % prior to purification, using the QIAquick pcr purification kit (Qiagen, Germany). Amplicons were sequenced in both directions using Sanger’s sequencing (GATC Biotech, Konstanz, Germany).

Sequences were trimmed and aligned with the BioEdit Sequence Alignment Editor (^{Hall, 1999}) and compared with the accessions of V. dahliae in GenBank using the Nucleotide blast tool.

Pathogenicity test

To carry out the pathogenicity test, V. dahliae inoculum was grown from a young culture in pdao medium. Three discs of fungus colonized medium (6 mm of diameter) were placed in 100 mL of potato broth with 2 % dextrose, and incubated under agitation at 110 rpm at room temperature (27 °C) for one week, to favor the development of the fungus. To remove the mycelia formed and to obtain the spore suspension used as inoculum, the broth was filtered through four layers of sterile gauze. The inoculum obtained was adjusted to a concentration of 10₇ cfu/mL (^{Resende et al., 1995}). For the inoculation of the pathogen, 40 plants from three-month-old ICS-1 T. cacao seeds (^{Resende et al., 1994}, ¹⁹⁹⁵) grown on sterile soil contained in black polyethylene bags were used.

The soil was autoclaved at 121 °C and 15 lb/in² for two consecutive days, for 45 and 30 min (^{Ormeno- Orrillo & Davila, 1999}). Two inoculation methods were evaluated: a) via drench to the soil with 50 mL of the conidia suspension (10₇ cfu/mL) around the root of 20 plants (^{Resende et al., 1995}); b) puncture in the stem of 20 plants employing a sterile needle.

Wounds were made in four points on the stem up to a depth of 2 mm (two on the bud of the cotyledonal leaves and two on the bud of the first true leaves), inoculating 5 μL of conidia suspension into each localized wounds (^{Trapero, Diez, Rallo, Barranco, & Lopez-Escudero, 2013}). Plants were maintained in the greenhouse of the ict Phytopathology Laboratory at a temperature of 28 °C ±} 0.57 °C, at a relative humidity of 70 % ±} 5.71 % and soil conditions at field capacity (^{Resende et al., 1995}).

On the other hand, the severity of the disease was evaluated with a scale of 0 to 6 (^{Villarroel- Zeballos, Feng, Iglesias, du Toit, & Correll, 2012}), where 0 = no symptoms; 1 = between 0 and 10 % of chlorosis or necrosis in the plant; 2 = between 10 and 25 % of chlorosis or necrosis; 3 = between 25 and 50 % chlorosis or necrosis; 4 = between 50 and 75 % of chlorosis or necrosis; 5 = between 75 and 100 % chlorosis or necrosis, and 6 = dead plant. Pathogen reevaluation was performed from sections of plant tissue sections (root and stem) in which the symptoms of the disease were observed, following the same isolation methodology described previously.

Endophytic Trichoderma (ET)

Endophytic Trichoderma (ET) isolates used in this study were obtained from the ict collection. Ten et isolates (table 1) with endophytic (^{Arevalo et al., 2010b}) and antagonistic capacities against Moniliophthora perniciosa, M. roreri and Phytophthora palmivora (^{Arevalo et al., 2010a}; ^{Leon et al., 2010a}, ^2010b), were selected to establish the antagonistic capacity against V. dahliae.

Table 1. Origin of endophytic Trichoderma (ET) isolates obtained from the stem and leaves of native cacao genotypes from localities of the Alto Amazonas region in Peru during 2008 

Source: Elaborated by the authors

Antibiosis

Soluble inhibitory metabolites were produced in three culture broths and the antibiosis test was adapted to the method described by Bailey et al. (²⁰⁰⁸). Trichoderma isolates were cultured in Petri dishes of 90 cm in diameter, containing malt extract agar (mea) culture medium and incubated at 25 °C +} 0 °C for five days. From these cultures a suspension of conidia was obtained by addition of sterile distilled water (sdw) to each plate. For this test, three culture broths were used: 1) based on minimum salts (min), 2) malt extract (me), and 3) potato sucrose (ps). For each broth, 1 mL of conidia suspension (106 cfu/mL) was added per 100 mL; then, it was incubated at 25 °C under agitation at 110 rpm for one week. Mycelia were filtered through a layer of sterile cotton and then centrifuged for 20 minutes at 2,500 rpm. The supernatant was stored at -20 °C until it was used.

Before being incorporated into the medium, the centrifugated filtrate was placed in a water bath at 90 °C +} 0 °C for two hours to inactivate the enzymes; then, an equal volume of psa medium with solidified agar (3 %) was added to the Petri dishes. A 4-mm-diameter broth slice of medium with the developed pathogen from a five-day-old colony was placed in the center of the Petri dish. Inhibition of the mycelial growth of the pathogen was recorded when the control covered the entire box (^{Osorio Hernandez et al., 2016}).

Mycoparasitism

For this test the precolonized plaque method described by Bailey et al. (²⁰⁰⁸) was used. A 2.5 × 0.5 cm piece of inoculum was extracted from a young culture of each Trichoderma isolate and located at one end of the plates containing pdao medium precolonized by V. dahliae. Three plates were used as replicates. The Petri dishes were incubated at 25 °C under darkness conditions.

Two weeks later the fungi were re-established. A total of eight samples (medium disks) of each box or replica were removed with a 5 mm diameter punch, from the point where the pathogen was planted. These samples were located in Petri dishes with pdao medium and incubated at 25 °C under darkness conditions; after 14 days the growth of Trichoderma and of the pathogen was observed, and the colonization percentage of Trichoderma biocontrol on medium disks with the pathogen was established.

Statistical análisis

Antibiosis and mycoparasitism tests were established under a completely randomized experimental design. Ten treatments were evaluated and corresponded to selected endophytic Trichoderma isolates. Three replicates per treatment were performed. The parameters evaluated were inhibition percentage of mycelial growth and colonization by Trichoderma of mycelium precolonized by V. dahliae. The percentage data was transformed angularly (arcsine √X) to confirm normality and homogeneity of variances as recommended by Montgomery (²⁰⁰⁸). An analysis of variance (andeva) and the Duncan contrast test with a level of significance of 0.05 % were carried out in the statistical software InfoStat version 2008 (^{Di Rienzo et al., 2008}).

In situ symptoms

Figure 1 shows the different symptoms observed in sudden death in a two-year-old CCN-51 T. cacao plantation located in the Atusparia sector, district and province of Tocache, Peru. Initially, flaccidity, epinasty, chlorosis, and necrosis of upper leaves are observed, and finally, wilting of the entire plant. When a cross-section of the upper limb was performed, necrotic points of the vascular system were identified and, in a longitudinal section, necrotic striae showing blackened infected vessels in the wood were observed. Similarly, necrotic rings were detected in the cross-section of the root of affected plants.

Figure 1. Symptoms of sudden death of cocoa in a plant of T. cacao CCN-51 clone located in the Atusparia sector, district and province of Tocache, Peru a. Yellowing, epinasty and wilting; b. Discoloration of the vascular system; c. Necrotic points throughout the vascular system; d. Necrotic striae in a longitudinal cut of the branch; and e. Necrotic rings in a cross section of the root. 

Morphological and molecular identification of Verticillium

The coloration of the colony in pdao medium was grayish to blackish due to microsclerotia formation (figure 2a). In the mycelium immersed in the middle and spase, with the help of a stereoscope, conidia immersed in mucilage were clearly observed as small water droplets. These colony characteristics correspond to those observed in the genus Verticillium (^{Barnett & Hunter, 1998}; ^{Barron, 1968}).

The conidiophores are hyaline, erect and carry a mass of conidia apically on each verticilated phialide in 2-3 fertile portions (figure 2b).

Figure 2. Macroscopic colony and microscopic morphology of vegetative and reproductive structures of Verticillium dahliae Kleb. a. The 15-days-old colony in potato dextrose agar (pda) medium; b. Conidia and verticilial phialides (40 X); c. Microsclerosis (40X). 

Conidia are hyaline, smooth-walled, ellipsoidal, unicellular, non-septated, and measure on average 6.7 ±} 1.4 μm × 2.9 μm ±} 0.5 μm. In figures 2a and 2c, microesclerotes immersed in the pdao culture medium regularly distributed throughout the colony, are blackish brown in color, in chains and aggregated, rounded and with a thin cell wall. The microscopic morphology characteristics mentioned place the pathogen as belonging to the species V. dahliae (^{Inderbitzin et al., 2011}; ^{Watanabe, 2010}).

Blast sequence analysis revealed that the Verticillium isolate showed 100 % homology to the V. dahliae sequences in GenBank (accession number NCBI: NR_126124.1 and HQ206718.1) and V. longisporum (HQ206832.1).

Pathogenicity

With the two methods of inoculation the Koch postulates were achieved and the causal agent of the sudden death of T. cacao was confirmed to be the species V. dahliae. Both the drench to the soil and the stem puncture inoculation methods using suspensions of V. dahliae conidia were able to reproduce wilting symptoms in 28 young inoculated cacao plants. However, the incubation period in the plants inoculated using the drench to the soil method was longer, and in which the first symptoms of leaf flaccidity followed by chlorosis and necrosis of the upper leaf apex were observed in ten plants (50 %) 30 days after inoculation (dai) with the pathogen (figure 3a); furthermore, ten days later the first symptoms of necrosis in the leaf area of the upper leaves (figure 3b) with a severity of 55 % ±} 1.70 were perceived. However, with the stem puncture method, the first symptoms were observed 10 dai (figure 3c) in 18 plants (90 % ±} 2.37), showing symptoms similar to those obtained with the drench to the soil method. The progression of the disease was severe; four days after observing the first symptoms, epithelia, wrinkling and necrosis of all leaves of the plant (figure 3d) occurred, and two days after that, some plants died showing a severity of 80 % ±} 2.38 % (table 2).

Table 2 Incidence and severity of sudden death of T. cacao by V. dahliae employing two inoculation methods at the Juan Bernito experimental station of ict, district of La Banda de Shilcayo, province and region of San Martin, Peru, in 2010 

Source: Elaborated by the authors

Note: Average data of 20 plants for severity per inoculation method. Different letters indicate significant differences according to the Duncan test (p ≤ 0.05).

Figura 3. Symptoms of pathogenicity with V. dahliae in plants of T. cacao ICS-1 clone employing the drench to the soil inoculation method. a. Wilt; b. Necrosis caused by the stem puncture inoculation method; c. Chlorosis; d. Wilt; e. Healthy plant; f. Stem showing necrotic striae; g and h. Necrotic points in the vascular system; i. Pathogen observed in pdao medium (inside the yellow circles). 

The current study comprises the first report that shows the presence of the sudden death of cacao plants in Peru, corroborating with the Koch´s postulates that V. dahliae is the causal agent of the disease. Of the 40 plants inoculated with the pathogen, 50 % and 90 % showed typical symptoms of the disease with the drench to the soil and the stem puncture inoculation methods, respectively. This disease caused by V. dahliae was reported in Uganda in 1965 (^{Emechebe, Leakey, & Banage, 1971}; ^{Leakey, 1965}), and in Brazil it has became a serious problem in the states of Bahia and Espiritu Santo, where 85 % of the cocoa is produced (^{Resende et al., 1995}). In addition, T. cacao in north-eastern Brazil was considered as a favorite host for Verticillium, because all Brazilian isolates of T. cacao were reported as pathogenic and destructive (^{Resende et al., 1994}).

In this study, V. dahliae infections were studied in three-month-old cacao plants of the ICS-1 clone using the drench to the soil and the stem puncture methods, finding that the first symptoms were observed 30 days with the first method. On the contrary, Resende et al. (¹⁹⁹⁴) reported the first symptoms at an earlier development stage, i.e. after only 20 days, and after 36 days, the disease rate was higher causing wilt in a small number of ICS-1 clone plants. Probably, this difference in days is due to the climatic conditions found in each location, as the temperature (28 °C ±} 0.57 °C) and the relative humidity (70 % ±} 5.71 %) were different from those reported by Resende et al. (¹⁹⁹⁴) (temperature of 25 °C ±} 2 °C and relative humidity of 60 % ±} 10 %).

However, the progression of the disease was rapid and showed more severe symptoms in 90 % of the seedlings inoculated by stem puncture, which agrees with what Resende et al. (¹⁹⁹⁵) reported.

These last authors used the same inoculation method, inoculum concentration and cacao plant genotypes (ICS-1 clone), observing the first symptoms after 10 days, which probably shows that the isolated pathogen of Peru has the same behavior as the one from Brazil. In addition, with the re-isolation of the pathogen, Koch's postulates were fulfilled and internal distribution within the plant was established.

Likewise, Trapero et al. (²⁰¹³) indicated that inoculation through stem injection in young olive seedlings (40 days old) was the most effective inoculation method to begin the selection process of olive genotypes resistant to the wilting due to Verticillium, constituting an effective method that saves work, space and resources. This method was also used to isolate cacao V. dahliae in Abelmoschus esculentus Medik and Gossypium hirsutum L. (Malvaceae), which induced similar symptoms on artificially inoculated T. cacao seedlings (^{Emechebe et al., 1971}).

Antagonistic capacity of endophytic Trichoderma against V. dahlia

In the antibiosis test, the endophytic Trichoderma cacao isolates were different in their ability to prevent the mycelial growth of V. dahliae in psa medium with mixtures of inhibitory metabolites (table 3).

Table 3 Effect of inhibitory metabolites of endophytic Trichoderma isolates obtained in three culture broths on the percentage of mycelial inhibition of V. dahliae. 

Source: Elaborated by the authors

Note: min: Broth based on mineral salts; ps: potato sucrose broth; me: Malt extract broth. Averages followed by different letters indicate significant differences between treatments according to the Duncan test (p ≤ 0.05). Values to the right of the mean indicate the standard error.

The metabolites obtained in ps medium had a higher effect on the mycelial inhibition (29.1 %) of V. dahliae, followed by min (22.2 %) compared to the metabolites obtained in me medium, which showed the lowest percentage (16.6 %). According to the mean culture media data, metabolites of the isolates TE-91, TE-17 and TE-126 prevented the growth of the pathogen with 66.5 %, 55.2 % and 51.9 %, respectively, compared to the other isolates that inhibited less than 17.2 %. Figure 4 shows that metabolites obtained from the TE-91 isolate in min broth prevented completely the mycelial growth of V. dahliae, followed by TE-17 with 77.2 % and 75.0 % in ps and me media, respectively; additionally, TE-126 with 72.8 % and 70.5 % in ps and me media, respectively, and TE-22 with 44.3 % in min. The other isolates had a low percentage of inhibition, i.e. less than 13.5 %.

Figure 4. Inhibition of mycelial growth of V. dahliae with soluble inhibitory metabolites of endophytic Trichoderma. 

In the myoparasitic test, the anova showed significant differences between the endophytic Trichoderma isolates (p = 0.05) in the colonization of V. dahliae (figure 5). The isolates TE-91 and TE- 126 were the most aggressive mycoparasites of V. dahliae with 100 % colonization on the pathogen, followed by TE-68, TE-72 and TE-76 with 80 %, 70 % and 87 % colonization, respectively, compared to isolates TE-17, TE-22, TE-50 and TE-55, which colonized V. dahliae in percentages lower than 30 %.

Source: Elaborated by the authors

Figure 5. Percentage of endophytic Trichoderma colonization on V. dahliae. Different letters indicate significant differences according to the Duncan test (p ≤ 0.05). 

Antibiosis, the production of antimicrobial compounds, mycoparasitism and the feeding of a fungus by another organism are some of the mechanisms of Trichoderma species that provide protection to plants against pathogens ( ^{Junaid et al., 2013}; ^{Martinez, Infante , & Reyes, 2013}). The colonization and production of soluble inhibitory metabolites of endophytic Trichoderma isolates obtained from different media varied in their ability to mycoparasite and inhibit the mycelial growth of V. dahliae in cccc medium, respectively.

This agrees with the results reported by Bailey et al. (²⁰⁰⁸), whose Trichoderma isolates varied greatly in their ability to produce inhibitory compounds of M. roreri and in their ability to parasitize these. In this study, Trichoderma asperellum (TE-91) proved to be the most aggressive mycoparasite that colonized and completely inhibited mycelial growth of the pathogen. This is probably due to a combined action between secondary metabolites such as terpenoids, which this species of Trichoderma releases into the pathogen, that inhibits germination and the synergistic interactions of hydrolytic enzymes such as chitinases, glucanases and proteases that degrade the cell wall of the pathogen until it causes its death (^{Lorito, Farkas, Rebuffat, Bodo, & Kubicek, 1996}; ^{Sivasithamparam & Ghisalberti, 1998}; ^{Zeilinger, Gruber, Bansal, & Mukherjee, 2016}). In addition, isolates of Trichoderma species have been reported as biocontrollers of V. dahliae (^{Carrero-Carron et al., 2016}; ^{Gao, Jun, Pan, Wu, & Zhang, 2007}; ^{Jabnoun- Khiareddine, Daami-Remadi, Ayed, Mahjoub, 2009}; ^{Meng, Tang, Huang, Ye, & Kan, 2007}) and Phytophthora megacarya (^{Mbarga et al., 2014}).

Furthermore, culture broths have an influence on the production of inhibitory metabolites such as non-ribosomal peptides (peptabiotics, siderophores and diketopiperazines), such as gliotoxin and gliovirine, polyketides, terpenes, pyrones and isocyano metabolites, which vary with the species of fungi and the environmental conditions (^{Zeilinger et al., 2016}). The metabolites of isolates TE-17, TE-91 and TE-126 obtained in ps culture were the ones that inhibited most of the mycelial growth of V. dahliae and belong to the species T. brevicompactum, T. asperellum and T. atroviride, respectively. At the same time, TE-126 also proved to be a very aggressive mycoparasite, to such an extent that the pathogen could not be recovered.

Trichoderma atroviride, reported as a mycoparasite, responds to a number of external stimuli, produces hydrolytic enzymes in the presence of a fungal host and wraps around the host hyphae (^{Harman, 2006}). In the same way, Degenkolb, Grafenhan, Nirenberg, Gams and Bruckner (²⁰⁰⁶) analyzed the formation of polypeptide antibiotics from four isolates of T. brevicompactum selected for their potential antagonism against regressive death. Isolates TE-50, TE-55, TE-68 and TE-76 belonging to the T. harzianum species inhibited the mycelial growth of the pathogen in the antibiosis test. However, TE-68 and TE-76 turned out to be aggressive mycoparasites. Bailey et al. (²⁰⁰⁸) indicated that T. harzianum isolates tend to have moderate to low antibiosis activity, and moderate to high mycoparasitism activity. This is probably due to the parallel formation of hydrolytic enzymes, together with a group of membrane-activating polypeptides called peptabiotics, whose synergistic action induces an important role in mycoparasitism between T. harzianum and its fungal host (^{Lorito et al. 1996}; ^{Schirmbock et al., 1994}).

The selection of Trichoderma strains with good endophilic capacity, in addition to other attributes such as mycoparisitism, antibiosis or induced resistance, can greatly improve the possibilities of developing strategies for biological control of cacao diseases (^{Bailey et al., 2008}).